Repairing An Old Barometer
07/07/2014
I have an old barometer which is a family hand-me-down.
Wikipedia tells me it is an aneroid type. Presumably, because there is no maker's mark, it is a relatively cheap mass-produced model. It is marked "MADE IN ENGLAND" which would date it roughly to before the 1960s, when "MADE IN JAPAN" began to make its appearance, and my school used to have Mitsubishi pencils. I recall that when I was a child here in New Zealand, once a British colony, perhaps 80% of all manufactured goods we saw were "Made In England", perhaps 15% were NZ made, and occasionally we saw "Made In USA". The barometer is not much use these days, as it is calibrated in Inches (of mercury), and those units probably went out when New Zealand went metric in the early Seventies. I could probably still sell it as an antique on the USA market! |
Nevertheless, when I found the indicating needle had detached from the shaft that drives it, I decided to make a project of repairing and perhaps re-calibrating the instrument.
Initial dismantling by removing the bezel which retains the glass cover revealed that the needle had come away from the hub used to mount the needle on the shaft.
The hub was still attached to the shaft. The hub is bigger than the hole in the face; to work on reattaching the needle to the hub, and also to access the mechanism behind, the hub must be gotten off the shaft somehow.
Searching online revealed that the hub should just be a friction fit on the shaft. Hoping to avoid bending or breaking the shaft, I used two small screwdrivers, one on each side of the hub to bravely apply force until, fortunately, it popped off without damage.
Initial dismantling by removing the bezel which retains the glass cover revealed that the needle had come away from the hub used to mount the needle on the shaft.
The hub was still attached to the shaft. The hub is bigger than the hole in the face; to work on reattaching the needle to the hub, and also to access the mechanism behind, the hub must be gotten off the shaft somehow.
Searching online revealed that the hub should just be a friction fit on the shaft. Hoping to avoid bending or breaking the shaft, I used two small screwdrivers, one on each side of the hub to bravely apply force until, fortunately, it popped off without damage.
The picture at left shows the kind of mechanism revealed once the mechanism case had been unscrewed from its recess in the back of the wooden mount, and the metal cover was removed from the case.
You can see an off-set circular metal bellows with a vacuum inside. The bellows responds to changes in air pressure by expanding and conracting. The mechanism works by having a series of links and levers convert and magnify the expansion and contraction of the bellows into a rotation of the needle's shaft. |
The needle was a little out of shape. I flattened it by gently hammering it aginst the back of a steel rule using the handle of a pair of pliers, which was nicely rounded and covered with plastic firm enough to trasmit force, yet not hard enough to mark or damage the needle.
I then found the hole in the needle for the hub was was ever so slightly too small. I very cautiously drilled it out to fit over the hub.
The needle was then attached to the hub with a tiny amount of oh-so-modern superglue.
In the centre of the outside of the glass cover is a small knurled knob attached to a small needle on the iside of the cover. It is not attached to anything else, and should be free to rotate by hand to mark and keep the place of the last reading, thus revealing how much the needle has moved since it was set. This was pretty stiff; I cleaned it and lubricated it slightly, then cleaned the glass inside and out.
I then put the mechanism back inside the wooden mount, with the shaft extending through to the front of the face. By consulting the NZ Meteorological Service's website for the current pressure, I put the needle and hub on the shaft near enough to the right reading.
I then reassembled everything else, and using the adjustment screw on the back of the mechanism's case, readjusted the needle to read exactly as per the current Met. Service value.
All that remained was recalibration. I dedicatedly took readings every day and entered them to a spreadsheet with date, time, barometer reading in inches, a conversion to hectopascals, and the Met. Service readings for my location. The resulting graph of the readings is shown below.
It is interesting to see that initially, up to around 04/07 the readings were accurately following the Met. Service ups and downs, but not reaching as high or low. Then I have a note in my spreadsheet file, "did more tapping", which is a reference to something I used to see my father do. When taking a reading, a couple of light taps on the knurled knob in the centre of the glass will vibrate the mechanism enough to overcome any slight friction, and enable the pointer to settle a bit further in whatever direction it is trending. After that the readings follow the Met Service a bit more closely. The error is about 1.5 Hectopascals over a range of 30, which is 5%.
p.s. I did find out later that one of the many little screws visible in the machanism can adjust the gain, i.e. magnifies the difference between the high and low readings. This would bring the readings even closer to the Met. Service, as they are about equally too low on the highs, and too high on the lows. However, I didn't worry about taking it apart again to try this adjustment.
I then found the hole in the needle for the hub was was ever so slightly too small. I very cautiously drilled it out to fit over the hub.
The needle was then attached to the hub with a tiny amount of oh-so-modern superglue.
In the centre of the outside of the glass cover is a small knurled knob attached to a small needle on the iside of the cover. It is not attached to anything else, and should be free to rotate by hand to mark and keep the place of the last reading, thus revealing how much the needle has moved since it was set. This was pretty stiff; I cleaned it and lubricated it slightly, then cleaned the glass inside and out.
I then put the mechanism back inside the wooden mount, with the shaft extending through to the front of the face. By consulting the NZ Meteorological Service's website for the current pressure, I put the needle and hub on the shaft near enough to the right reading.
I then reassembled everything else, and using the adjustment screw on the back of the mechanism's case, readjusted the needle to read exactly as per the current Met. Service value.
All that remained was recalibration. I dedicatedly took readings every day and entered them to a spreadsheet with date, time, barometer reading in inches, a conversion to hectopascals, and the Met. Service readings for my location. The resulting graph of the readings is shown below.
It is interesting to see that initially, up to around 04/07 the readings were accurately following the Met. Service ups and downs, but not reaching as high or low. Then I have a note in my spreadsheet file, "did more tapping", which is a reference to something I used to see my father do. When taking a reading, a couple of light taps on the knurled knob in the centre of the glass will vibrate the mechanism enough to overcome any slight friction, and enable the pointer to settle a bit further in whatever direction it is trending. After that the readings follow the Met Service a bit more closely. The error is about 1.5 Hectopascals over a range of 30, which is 5%.
p.s. I did find out later that one of the many little screws visible in the machanism can adjust the gain, i.e. magnifies the difference between the high and low readings. This would bring the readings even closer to the Met. Service, as they are about equally too low on the highs, and too high on the lows. However, I didn't worry about taking it apart again to try this adjustment.